基于机器学习与无人机高光谱的九龙江口水质参数定量反演OA
Quantitative inversion of water quality parameters in the Jiulong River Estuary based on machine learning and UAV hyperspectral data
[背景]机器学习结合无人机(UAV)高光谱遥感在水质动态精细监测方面具有巨大潜力,现有研究多见于湖河水库等稳定水体,对于自然过程与人类活动耦合干扰强烈的河口区域,水体光学特性复杂多变,相关研究成果较为有限.[目的]验证该技术应用于河口复杂环境的有效性和可行性.[方法]本研究在九龙江口开展了多潮时连续的无人机高光谱测量同步水样采集实验,通过不同特征筛选策略[Pearson 相关性分析、主成分分析(PCA)、连续投影算法(SPA)]结合机器学习算法[支持向量回归(SVR)、随机森林(RF)、多层感知机(MLP)]构建叶绿素a 和悬浮泥沙反演模型,评估不同建模方案的反演效果.[结果]结果显示,叶绿素a 最优反演模型为Pearson-RF,决定系数(R2)为0.83,均方根误差(RMSE)为 0.41 μg/L,相对分析误差(RPD)为 2.50;悬浮泥沙最优反演模型为 Pearson-RF,R2为 0.90,RMSE 为 3.63 mg/L,RPD 为 3.26,PCA-MLP 亦表现出可靠的预测性能(R2=0.89,RMSE=3.85 mg/L,RPD=3.07).[结论]通过本研究,发现 Pearson-RF 模型特征提取和抗干扰能力优异,兼具良好的反演精度和泛化性能,在小样本、弱信号场景的应用优势明显,适合作为叶绿素 a 和悬浮泥沙反演的通用算法.其中,对于光谱响应特征突出的悬浮泥沙,PCA-MLP 模型也可实现较高精度预测.研究成果可为河口水质定量反演、动态监测和精细化管理提供方法参考和技术支撑.
[Background]The combination of machine learning and unmanned aerial vehicle(UAV)hyper-spectral remote sensing demonstrates great potential for dynamic and refined water quality monitoring.Most ex-isting studies focus on stable water bodies such as lakes,rivers and reservoirs,while relevant research remains relatively scarce for estuaries with complex optical properties,affected by intense coupled disturbances from natural processes and human activities.[Objective]The study aims to verify the effectiveness and feasibility of the proposed technique for application in estuarine regions.[Methods]This study conducted continuous UAV hyperspectral measurements and simultaneous water sampling across multiple tidal phases in the Jiulong River Estuary.Building on this,quantitative inversion models for chlorophyll-a and suspended sediment were de-veloped and evaluated by combining different feature selection methods—Pearson correlation analysis,princi-pal component analysis(PCA),and the successive projections algorithm(SPA)—with machine learning al-gorithms—support vector regression(SVR),random forest(RF),and multilayer perceptron(MLP).[Results]The results indicated that the optimal inversion model for chlorophyll-a was Pearson-RF,with coefficient of de-termination(R2)of 0.83,root mean square error(RMSE)of 0.41 μg/L,and residual predicitive deviation(RPD)of 2.50.The optimal model for suspended sediment was Pearson-RF,with R2 of 0.90,RMSE of 3.63 mg/L and RPD of 3.26.The PCA-MLP model also showed reliable predictive performance in the retrieval of suspended sediment(R2=0.89,RMSE=3.85 mg/L,RPD=3.07).[Conclusion]Through this study,the Pearson-RF model exhibits excellent feature extraction capability and robustness,along with favorable inversion accuracy and gen-eralization performance.In practical applications,the Pearson-RF model demonstrates prominent advantages in the small-sample and weak-signal scenarios,rendering it a universal algorithm suitable for the inversion of chlorophyll-a and suspended sediment concentrations.Specifically,the PCA-MLP model can achieve reason-able prediction accuracy for suspended sediment with distinct spectral response characteristics.These findings can provide methodological reference and technical support for quantitative inversion,dynamic monitoring and refined management of water quality in estuarine regions.
陈嘉伟;俞俊毅;蔚广鑫;骆巧琦
福建海洋研究所,福建 厦门 361013||福建省海岛与海岸带管理技术研究重点实验室,福建 厦门 361013福建海洋研究所,福建 厦门 361013福建海洋研究所,福建 厦门 361013福建海洋研究所,福建 厦门 361013
农业科技
机器学习无人机(UAV)高光谱九龙江口叶绿素a悬浮泥沙特征筛选定量反演
machine learningunmanned aerial vehicle(UAV)hyperspectral datathe Jiulong River Estuarychlorophyll-asuspended sedimentfeature selectionquantitative inversion
《渔业研究》 2026 (5)
590-604,15
福建省属公益类科研院所基本科研专项(2021R1007001、2023R1007002)
评论